摘要
为了有效地对不同深度的局部腐蚀缺陷超声波信号进行分类识别,根据腐蚀缺陷信号样本数量较少的特点,提出了一种基于主成分分析(PCA)和支持向量机(SVM)的超声波腐蚀缺陷信号识别方法。该方法采用经验模态分解法对腐蚀缺陷信号进行分解,提取各本征模式分量的时域无量纲参数,利用主成分分析消除原始特征集中的冗余信息,降低每一个特征之间的相关性,实现腐蚀缺陷信号特征参数的降维。在PCA进行特征优化后,将支持向量机的多类分类应用于缺陷分类过程中。将腐蚀缺陷原始特征集和经主成分分析优化后的特征集,分别用于支持向量机的训练和测试,且选择不同的核函数构造支持向量机分类器。实验结果表明,基于主成分分析和支持向量机的方法可以有效地对超声波腐蚀缺陷深度信号分类。
In order to effectively identify and classify in different depth of local corrosion defect by ultrasonic signal classification,according to the characteristics of corrosion defect signals of the small number of samples,propose a method based on principal component analysis( PCA) and support vector machine( SVM) ultrasonic corrosion defect signal recognition method. The method uses empirical mode decomposition to decompose the signal of corrosion defect,extracts the non dimensional parameters of each eigenmode component in time domain,and uses Principal component analysis to eliminate redundant information in original feature set,reduce correlation between each feature,and achieve dimension reduction of corrosion defect signal feature parameters.After optimizing the features of PCA,the multi-class classification of support vector machines is applied to the process of defect classification. The original feature set of corrosion defects and the optimized feature set after principal component analysis are used for training and testing of support vector machines,respectively,and support vector machine classifier is constructed by selecting different kernel functions. The experimental results show that the method based on principal component analysis and support vector machine can effectively classify the ultrasonic corrosion defect signals.
引文
[1]郭华杰.管道超声导波检测信号的分析与识别评估研究[D].江苏大学,2015.
[2]蔡敏.海底输油管道超声检测信号的分析与缺陷评估的研究[D].上海:上海交通大学,2008.
[3]戴波,赵晶,周炎.超声波管道内检测腐蚀缺陷分类识别研究[J].机床与液压,2008,36(b07):194-198.
[4]周西峰,索会迎,郭前岗,等.基于小波包-BP网络的超声检测缺陷类型识别[J].解放军理工大学自然科学版,2012,13(5):521-526.
[5]Sambath S,Selvakumar N.Automatic Defect Classification in Ultrasonic NDT Using Artificial Intelligence[J].Journal of Nondestructive Evaluation,2011,30(1):20-28.
[6]Liu J,Xu G,Ren L,et al.Defect intelligent identification in resistance spot welding ultrasonic detection based on wavelet packet and neural network[J].International Journal of Advanced Manufacturing Technology,2016:1-8.
[7]万振凯,王占刚.基于神经网络的复合材料缺陷超声波检测研究[J].纺织学报,2010,31(2):54-59.
[8]Garcia-Alvarez D,Fuente M J,Sainz G I.Fault Detection and Isolation in Transient States Using Principal Component Analysis[J].Journal of Process Control,2012,22(3):551-563.
[9]He Q,Yan R,Kong F,Du R,et al.Machine Condition Monitoring Using Principal Component Representations[J].Mechanical Systems and Signal Processing,2009,23(2):446-466.
[10]Subbaraj P,Kannapiran B.Fault Diagnosis of Pneumatic Valve Using PCA and ANN Techniques[J].Communications in Computer and Information Science,2011,204:404-413.
[11]刘永斌,何清波,孔凡让,等.基于PCA和SVM的内燃机故障诊断[J].振动、测试与诊断,2012,32(2):250-255.
[12]皮子坤,贾宝山,贾廷贵,等.煤矿瓦斯涌出量动态预测的PCA-MFOA-GRNN模型及应用[J].传感技术学报,2015(11):1676-1681.
[13]Chang Yanwei,Wang Y C,Tao L,et al.Fault Diagnosis of a Mine Hoist Using PCA and SVM Techniques[J].International Journal of Mining Science and Technology(矿业科学技术学报),2008,18(3):327-331.
[14]Virupakshappa K,Oruklu E.Ultrasonic Flaw Detection Using Support Vector Machine Classification[C]//Ultrasonics Symposium.IEEE,2015:1-4.
[15]Tsao W C,Li Y F,Le D D,et al.An Insight Concept to Select Appropriate IMFs for Envelope Analysis of Bearing Fault Diagnosis[J].Measurement,2012,45(6):1489-1498.
[16]徐卓飞,刘凯,张海燕,等.基于经验模式分解和主元分析的滚动轴承故障诊断方法研究[J].振动与冲击,2014,33(23):133-139.
[17]Lei Y,He Z,Zi Y.Application of an Intelligent Classification Method to Mechanical Fault Diagnosis[J].Expert Systems with Applications,2009,36(6):9941-9948.